{"id":{"repo_id":"unf","oai_identifier":"oai:digitalcommons.unf.edu:etd-2467"},"canonical_url":"https://search.ndltd.org/etd/unf/oai:digitalcommons.unf.edu:etd-2467","repository":{"repo_id":"unf","name":"University of North Florida","base_url":"https://digitalcommons.unf.edu/do/oai/"},"display":{"title":"Solar energy prediction using advanced hybrid machine learning models","abstract":"Accurate short-term forecasting of solar power generation is critical for the reliable and cost-effective operation of renewable-based microgrids, where sudden weather-induced variability can compromise grid stability, battery scheduling, and energy trading decisions. Traditional physical and statistical models struggle to capture the complex non-linear relationships and localized weather effects, while individual deep learning architectures often exhibit systematic biases such as chronic under-prediction of peak generation. This thesis proposes a novel Cross-Feedback Ensemble framework that combines the complementary strengths of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and one-dimensional Convolutional Neural Network (1D-CNN) models through an iterative cross-feedback mechanism and a linear meta-learner. Unlike conventional stacking or error-gated ensembles, the proposed architecture allows each base model to receive the real-time predictions of the others as additional input features during retraining phases, enabling mutual error correction and richer representation of temporal, short-term, and spatial patterns in weather-to-power relationships. The models were developed and rigorously evaluated using four years (2016–2019) of hourly solar generation and meteorological data for Germany from the Open Power Systems Data (OPSD) platform. The proposed Cross-Feedback Ensemble achieved a Mean Absolute Error (MAE) of 512.3 MW, Mean Squared Error (MSE) of 883,954.54 MW², and coefficient of determination (R²) of 0.9756 on the hold-out test set — representing improvements of approximately 11.9%, 14%, and 0.41%, respectively, over a baseline stacking ensemble and even larger gains compared to standalone LSTM, GRU, and CNN models. Most notably, the cross-feedback mechanism virtually eliminated the systematic under-prediction bias observed in individual deep learning models. The results demonstrate that explicit inter-model feedback and dynamic error correction significantly enhance forecasting accuracy and robustness in real-world variable weather conditions. The proposed framework offers a practical, high-performance solution for microgrid energy management systems and can be extended to multi-source renewable forecasting and real-time operational deployment.","abstract_html":"Accurate short-term forecasting of solar power generation is critical for the reliable and cost-effective operation of renewable-based microgrids, where sudden weather-induced variability can compromise grid stability, battery scheduling, and energy trading decisions. Traditional physical and statistical models struggle to capture the complex non-linear relationships and localized weather effects, while individual deep learning architectures often exhibit systematic biases such as chronic under-prediction of peak generation. This thesis proposes a novel Cross-Feedback Ensemble framework that combines the complementary strengths of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and one-dimensional Convolutional Neural Network (1D-CNN) models through an iterative cross-feedback mechanism and a linear meta-learner. Unlike conventional stacking or error-gated ensembles, the proposed architecture allows each base model to receive the real-time predictions of the others as additional input features during retraining phases, enabling mutual error correction and richer representation of temporal, short-term, and spatial patterns in weather-to-power relationships. The models were developed and rigorously evaluated using four years (2016–2019) of hourly solar generation and meteorological data for Germany from the Open Power Systems Data (OPSD) platform. The proposed Cross-Feedback Ensemble achieved a Mean Absolute Error (MAE) of 512.3 MW, Mean Squared Error (MSE) of 883,954.54 MW², and coefficient of determination (R²) of 0.9756 on the hold-out test set — representing improvements of approximately 11.9%, 14%, and 0.41%, respectively, over a baseline stacking ensemble and even larger gains compared to standalone LSTM, GRU, and CNN models. Most notably, the cross-feedback mechanism virtually eliminated the systematic under-prediction bias observed in individual deep learning models. The results demonstrate that explicit inter-model feedback and dynamic error correction significantly enhance forecasting accuracy and robustness in real-world variable weather conditions. The proposed framework offers a practical, high-performance solution for microgrid energy management systems and can be extended to multi-source renewable forecasting and real-time operational deployment.","abstract_has_math":false,"creators":["Alqhtani, Jelawi A","Alqhtani, Jelawi, 1987-"],"institution":"University of North Florida. School of Engineering","degree_name":"Master of Science in Electrical Engineering (MSEE)","degree_level":null,"degree_discipline":"Engineering","degree_department":null,"school":null,"contributors":["Dr. Touria El Mezyani","Dr. Xudong Liu","Dr. Hemani Kaushal"],"advisors":["El Mezyani, Touria","Liu, Xudong","Kaushal, Hemani"],"committee_chairs":[],"committee_members":[],"year":null,"date_issued":"Fall","date_published":"Fall","updated_at":"2026-11-18T20:04:15Z","subjects":["Thesis","University of North Florida","UNF","Dissertations, Academic -- UNF -- Master of Science in Electrical Engineering","Dissertations, Academic -- UNF -- Engineering","machine learning, solar energy, prediction, artificial intelligence, ensemble, solar power forecasting, microgrid energy management, deep learning, ensemble methods, cross-feedback ensemble, LSTM, GRU, CNN, renewable energy integration, neural networks","Computational Engineering","Electrical and Computer Engineering","Electrical and Electronics","Other Computer Engineering","Other Electrical and Computer Engineering","Power and Energy","Solar energy -- Forecasting -- Technological innovations; Renewable energy sources -- Technological innovations; Machine learning -- Industrial applications -- Testing; Artificial intelligence -- Industrial applications -- Testing"],"languages":[],"rights":[],"rights_urls":["http://rightsstatements.org/vocab/InC/1.0/"],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"https://digitalcommons.unf.edu/etd/1374","outbound_label":"Repository record","outbound_source":"dc:identifier","record_url":"https://digitalcommons.unf.edu/etd/1374","record_label":"Repository record","record_kind":"record","record_source":"dc:identifier","download_url":"https://digitalcommons.unf.edu/context/etd/article/2467/viewcontent/ETD1949PA.pdf","download_label":"Download PDF","download_source":"dc:identifier","access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://digitalcommons.unf.edu/do/oai/?verb=GetRecord&metadataPrefix=oai_etdms&identifier=oai%3Adigitalcommons.unf.edu%3Aetd-2467","prefix":"oai_etdms"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Dr. Touria El Mezyani","Dr. Xudong Liu","Dr. Hemani Kaushal"]},{"key":"dc:contributor.advisor","label":"Advisor","values":["El Mezyani, Touria","Liu, Xudong","Kaushal, Hemani"]},{"key":"dc:creator","label":"Author","values":["Alqhtani, Jelawi A","Alqhtani, Jelawi, 1987-"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.available","label":"Dc Date Available","values":["2026-01-06T08:00:00Z"]},{"key":"dc:date.created","label":"Dc Date Created","values":["2025-01-01T08:00:00Z"]},{"key":"dc:date.issued","label":"Date","values":["Fall"]},{"key":"dc:type","label":"Dc Type","values":["Master's Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Engineering"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Master of Science in Electrical Engineering (MSEE)"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of North Florida. School of Engineering"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Thesis","University of North Florida","UNF","Dissertations, Academic -- UNF -- Master of Science in Electrical Engineering","Dissertations, Academic -- UNF -- Engineering","machine learning, solar energy, prediction, artificial intelligence, ensemble, solar power forecasting, microgrid energy management, deep learning, ensemble methods, cross-feedback ensemble, LSTM, GRU, CNN, renewable energy integration, neural networks","Computational Engineering","Electrical and Computer Engineering","Electrical and Electronics","Other Computer Engineering","Other Electrical and Computer Engineering","Power and Energy"]},{"key":"dc:subject.lcsh","label":"Dc Subject Lcsh","values":["Solar energy -- Forecasting -- Technological innovations; Renewable energy sources -- Technological innovations; Machine learning -- Industrial applications -- Testing; Artificial intelligence -- Industrial applications -- Testing"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:rights","label":"Dc Rights","values":["http://rightsstatements.org/vocab/InC/1.0/"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://digitalcommons.unf.edu/etd/1374","https://digitalcommons.unf.edu/context/etd/article/2467/viewcontent/ETD1949PA.pdf"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Accurate short-term forecasting of solar power generation is critical for the reliable and cost-effective operation of renewable-based microgrids, where sudden weather-induced variability can compromise grid stability, battery scheduling, and energy trading decisions. Traditional physical and statistical models struggle to capture the complex non-linear relationships and localized weather effects, while individual deep learning architectures often exhibit systematic biases such as chronic under-prediction of peak generation. This thesis proposes a novel Cross-Feedback Ensemble framework that combines the complementary strengths of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and one-dimensional Convolutional Neural Network (1D-CNN) models through an iterative cross-feedback mechanism and a linear meta-learner. Unlike conventional stacking or error-gated ensembles, the proposed architecture allows each base model to receive the real-time predictions of the others as additional input features during retraining phases, enabling mutual error correction and richer representation of temporal, short-term, and spatial patterns in weather-to-power relationships. The models were developed and rigorously evaluated using four years (2016–2019) of hourly solar generation and meteorological data for Germany from the Open Power Systems Data (OPSD) platform. The proposed Cross-Feedback Ensemble achieved a Mean Absolute Error (MAE) of 512.3 MW, Mean Squared Error (MSE) of 883,954.54 MW², and coefficient of determination (R²) of 0.9756 on the hold-out test set — representing improvements of approximately 11.9%, 14%, and 0.41%, respectively, over a baseline stacking ensemble and even larger gains compared to standalone LSTM, GRU, and CNN models. Most notably, the cross-feedback mechanism virtually eliminated the systematic under-prediction bias observed in individual deep learning models. The results demonstrate that explicit inter-model feedback and dynamic error correction significantly enhance forecasting accuracy and robustness in real-world variable weather conditions. The proposed framework offers a practical, high-performance solution for microgrid energy management systems and can be extended to multi-source renewable forecasting and real-time operational deployment."]},{"key":"dc:description.note","label":"Dc Description Note","values":["ProQuest: Yes"]},{"key":"dc:title","label":"Title","values":["Solar energy prediction using advanced hybrid machine learning models"]}]}],"canonical_facts":{"dc:contributor":["Dr. Touria El Mezyani","Dr. Xudong Liu","Dr. Hemani Kaushal"],"dc:contributor.advisor":["El Mezyani, Touria","Liu, Xudong","Kaushal, Hemani"],"dc:creator":["Alqhtani, Jelawi A","Alqhtani, Jelawi, 1987-"],"dc:date.available":["2026-01-06T08:00:00Z"],"dc:date.created":["2025-01-01T08:00:00Z"],"dc:date.issued":["Fall"],"dc:description.abstract":["Accurate short-term forecasting of solar power generation is critical for the reliable and cost-effective operation of renewable-based microgrids, where sudden weather-induced variability can compromise grid stability, battery scheduling, and energy trading decisions. Traditional physical and statistical models struggle to capture the complex non-linear relationships and localized weather effects, while individual deep learning architectures often exhibit systematic biases such as chronic under-prediction of peak generation. This thesis proposes a novel Cross-Feedback Ensemble framework that combines the complementary strengths of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and one-dimensional Convolutional Neural Network (1D-CNN) models through an iterative cross-feedback mechanism and a linear meta-learner. Unlike conventional stacking or error-gated ensembles, the proposed architecture allows each base model to receive the real-time predictions of the others as additional input features during retraining phases, enabling mutual error correction and richer representation of temporal, short-term, and spatial patterns in weather-to-power relationships. The models were developed and rigorously evaluated using four years (2016–2019) of hourly solar generation and meteorological data for Germany from the Open Power Systems Data (OPSD) platform. The proposed Cross-Feedback Ensemble achieved a Mean Absolute Error (MAE) of 512.3 MW, Mean Squared Error (MSE) of 883,954.54 MW², and coefficient of determination (R²) of 0.9756 on the hold-out test set — representing improvements of approximately 11.9%, 14%, and 0.41%, respectively, over a baseline stacking ensemble and even larger gains compared to standalone LSTM, GRU, and CNN models. Most notably, the cross-feedback mechanism virtually eliminated the systematic under-prediction bias observed in individual deep learning models. The results demonstrate that explicit inter-model feedback and dynamic error correction significantly enhance forecasting accuracy and robustness in real-world variable weather conditions. The proposed framework offers a practical, high-performance solution for microgrid energy management systems and can be extended to multi-source renewable forecasting and real-time operational deployment."],"dc:description.note":["ProQuest: Yes"],"dc:identifier":["https://digitalcommons.unf.edu/etd/1374","https://digitalcommons.unf.edu/context/etd/article/2467/viewcontent/ETD1949PA.pdf"],"dc:rights":["http://rightsstatements.org/vocab/InC/1.0/"],"dc:subject":["Thesis","University of North Florida","UNF","Dissertations, Academic -- UNF -- Master of Science in Electrical Engineering","Dissertations, Academic -- UNF -- Engineering","machine learning, solar energy, prediction, artificial intelligence, ensemble, solar power forecasting, microgrid energy management, deep learning, ensemble methods, cross-feedback ensemble, LSTM, GRU, CNN, renewable energy integration, neural networks","Computational Engineering","Electrical and Computer Engineering","Electrical and Electronics","Other Computer Engineering","Other Electrical and Computer Engineering","Power and Energy"],"dc:subject.lcsh":["Solar energy -- Forecasting -- Technological innovations; Renewable energy sources -- Technological innovations; Machine learning -- Industrial applications -- Testing; Artificial intelligence -- Industrial applications -- Testing"],"dc:title":["Solar energy prediction using advanced hybrid machine learning models"],"dc:type":["Master's Thesis"],"thesis:degree_discipline":["Engineering"],"thesis:degree_name":["Master of Science in Electrical Engineering (MSEE)"],"thesis:institution_name":["University of North Florida. School of Engineering"]},"updated_at":"2026-11-18T20:04:15Z"}